DM1-ML Supervised Learning

URL to have access to all the codes : https://github.com/ChristopheYe/DM1-ML.git

I. PROJECT'S TITLE

DM1 ML

II. PROJET DESCRIPTION

The code is written with Python on different Jupyter Notebook

Experimentation with five learning algorithms :

1. Decision trees with some form of pruning
2. k-nearest neighbors
3. Neural networks
4. Boosting
5. Support Vector Machines

All the learning algorithms used in this code come from the library sklearn.

All the different algorithms were tested on 2 different datasets :

1. Movie Dataset.csv
2. wine-quality-white-and-red.csv

In the first dataset, I want to know if a movie got an award or no In the send dataset, I want to know if the wine is a red wine or a white wine

For every learning algorithms, I always start with testing the function with its parameters set by default and make a cross validation on the test set. Then, I vary the hyperparameters with a validation and a learning curve and also look at the time needed for the program to execute. At the end, I see the difference in the results between the function with hyperparameters set by default and the one with hyperparameters chosen carefully.

III. HOW TO INSTALL AND RUN THE PROJECT

1. Download Anaconda-Navigator
2. Use a Jupyter Notebook